Generalized gradients and paths of descent

From MaRDI portal





The optimality condition used in subgradient analysis is that zero is a subgradient of f at x, i.e. x is a substationary point of a Lipschitz function \(f: R^ n\to R\). If x is not a substationary point, then there is a direction of uniform descent at x. The aim of the present paper is to show that if f is only lower semicontinuous and x is not a substationary point, then, even though the descent directions may not exist, there exist paths of descent. This is done using a characterization of Clarke's tangent cone in terms of Lipschitz paths.











This page was built for publication: Generalized gradients and paths of descent

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3719365)